<i>The reason why [Computer Science] research produces so little that can be called creative programming these days is that the modern process of grant-funded research is fundamentally incompatible with the task of writing interesting, cool and relevant software. Rather, its goal is to produce publications and careers, and it’s very good at that.</i><p>This is quite a naive point of view. To paraphrase Dijkstra:<p><i>Computer science is no more about programming than astronomy is about telescopes.</i><p>(s/programming/computers/ for the original)<p>Look at this list <a href="http://en.wikipedia.org/wiki/Computer_science#Fields_of_computer_science" rel="nofollow">http://en.wikipedia.org/wiki/Computer_science#Fields_of_comp...</a> and tell me what percentage of that consists of programming. It's like saying math is about numbers.
This is a sensitive topic for me, since I'm in love with studying.
I have to disagree with a few points, namely that universities have developed into bureaucracies primarily seeking to further careers. I worked as an undergraduate researcher in a neuroscience lab, and although publication was always a pressure, I don't know anyone who wasn't there because they didn't want to be. Studying the same subject on such an exclusive basis for 5 years or more can cause burnout, but everyone in the lab came in excited at the possibilities of solving new problems.
However, universities can do more to ensure that their work helps the community directly - big research needs to continue unabated, or perhaps even on an accelerated schedule, but students and many professors can shift their focus from class-centric papers and projects to community-based projects.
Actually, this is the startup I'm working on right now, and it came about because students "do not want to churn out meaningless solutions to irrelevant problems", they want to do meaningful things.
It was never the University's role to innovate in the manner that the article says they do. But just look at google summer of code: All of these projects—really, really cool things—are university research projects. Besides, it was never the university's role to innovate in the manner that the article insinuates: i.e. tangible code. Most of the "innovation", or creative code, you see in today's world are simply the application of already existing theories.
<i>The reason why [Computer Science] research produces so little that can be called creative programming these days...</i><p>I'd like to see some examples of "meaningless solutions to irrelevant problems". CS research was never really about producing "creative programming", but rather about furthering what we know about computers, which involves a lot of math and theoretical research.
I have a problem with the assumption that "publish or perish" is killing the universities. The assumption is that one has to publish a lot of material. That is not true. Gauss had an excellent quote that applies here, "few but ripe." This point behind the quote is publish a little, but let what you publish be truly brilliant. Do you honestly think Harvard's math department would not want the person who manages to prove the Riemann hypothesis if that person also only had two other published papers? The answer is resounding, "hell no." Academia is about quality, not quantity.
Given the number of people alive, and the number of research positions in universities, I have trouble accepting that this premise was <i>ever</i> true.